htslib 0.2.2 → 0.2.5

data/lib/hts/bam.rb

@@ -57,7 +57,6 @@ module HTS
       build_index(index) if build_index
       @idx = load_index(index)
       @start_position = tell
-      super # do nothing
     end
 
     def build_index(index_name = nil, min_shift: 0)
@@ -90,18 +89,21 @@ module HTS
       !@idx.null?
     end
 
-    # Close the current file.
     def close
       LibHTS.hts_idx_destroy(@idx) if @idx&.null?
       @idx = nil
       super
     end
 
+    def fai=(fai)
+      check_closed
+      LibHTS.hts_set_fai_filename(@hts_file, fai) > 0 || raise
+    end
+
     def write_header(header)
       check_closed
 
       @header = header.dup
-      LibHTS.hts_set_fai_filename(@hts_file, @file_name)
       LibHTS.sam_hdr_write(@hts_file, header)
     end
 
@@ -109,7 +111,8 @@ module HTS
       check_closed
 
       aln_dup = aln.dup
-      LibHTS.sam_write1(@hts_file, header, aln_dup) > 0 || raise
+      r = LibHTS.sam_write1(@hts_file, header, aln_dup)
+      raise "Failed to write record" if r < 0
     end
 
     def each(copy: false, &block)
@@ -120,29 +123,6 @@ module HTS
       end
     end
 
-    private def each_record_copy
-      check_closed
-      return to_enum(__method__) unless block_given?
-
-      while LibHTS.sam_read1(@hts_file, header, bam1 = LibHTS.bam_init1) != -1
-        record = Record.new(bam1, header)
-        yield record
-      end
-      self
-    end
-
-    private def each_record_reuse
-      check_closed
-      # Each does not always start at the beginning of the file.
-      # This is the common behavior of IO objects in Ruby.
-      return to_enum(__method__) unless block_given?
-
-      bam1 = LibHTS.bam_init1
-      record = Record.new(bam1, header)
-      yield record while LibHTS.sam_read1(@hts_file, header, bam1) != -1
-      self
-    end
-
     def query(region, copy: false, &block)
       if copy
         query_copy(region, &block)
@@ -151,45 +131,6 @@ module HTS
       end
     end
 
-    private def query_copy(region)
-      check_closed
-      raise "Index file is required to call the query method." unless index_loaded?
-      return to_enum(__method__, region) unless block_given?
-
-      qiter = LibHTS.sam_itr_querys(@idx, header, region)
-
-      begin
-        loop do
-          bam1 = LibHTS.bam_init1
-          slen = LibHTS.sam_itr_next(@hts_file, qiter, bam1)
-          break if slen == -1
-          raise if slen < -1
-
-          yield Record.new(bam1, header)
-        end
-      ensure
-        LibHTS.hts_itr_destroy(qiter)
-      end
-      self
-    end
-
-    private def query_reuse(region)
-      check_closed
-      raise "Index file is required to call the query method." unless index_loaded?
-      return to_enum(__method__, region) unless block_given?
-
-      qiter = LibHTS.sam_itr_querys(@idx, header, region)
-
-      bam1 = LibHTS.bam_init1
-      record = Record.new(bam1, header)
-      begin
-        yield record while LibHTS.sam_itr_next(@hts_file, qiter, bam1) > 0
-      ensure
-        LibHTS.hts_itr_destroy(qiter)
-      end
-      self
-    end
-
     # @!macro [attach] define_getter
     #   @method $1
     #   Get $1 array
@@ -247,5 +188,71 @@ module HTS
 
       self
     end
+
+    private
+
+    def query_reuse(region)
+      check_closed
+      raise "Index file is required to call the query method." unless index_loaded?
+      return to_enum(__method__, region) unless block_given?
+
+      qiter = LibHTS.sam_itr_querys(@idx, header, region)
+      raise "Failed to query region: #{region}" if qiter.null?
+
+      bam1 = LibHTS.bam_init1
+      record = Record.new(bam1, header)
+      begin
+        yield record while LibHTS.sam_itr_next(@hts_file, qiter, bam1) > 0
+      ensure
+        LibHTS.hts_itr_destroy(qiter)
+      end
+      self
+    end
+
+    def query_copy(region)
+      check_closed
+      raise "Index file is required to call the query method." unless index_loaded?
+      return to_enum(__method__, region) unless block_given?
+
+      qiter = LibHTS.sam_itr_querys(@idx, header, region)
+      raise "Failed to query region: #{region}" if qiter.null?
+
+      begin
+        loop do
+          bam1 = LibHTS.bam_init1
+          slen = LibHTS.sam_itr_next(@hts_file, qiter, bam1)
+          break if slen == -1
+          raise if slen < -1
+
+          yield Record.new(bam1, header)
+        end
+      ensure
+        LibHTS.hts_itr_destroy(qiter)
+      end
+      self
+    end
+
+    def each_record_reuse
+      check_closed
+      # Each does not always start at the beginning of the file.
+      # This is the common behavior of IO objects in Ruby.
+      return to_enum(__method__) unless block_given?
+
+      bam1 = LibHTS.bam_init1
+      record = Record.new(bam1, header)
+      yield record while LibHTS.sam_read1(@hts_file, header, bam1) != -1
+      self
+    end
+
+    def each_record_copy
+      check_closed
+      return to_enum(__method__) unless block_given?
+
+      while LibHTS.sam_read1(@hts_file, header, bam1 = LibHTS.bam_init1) != -1
+        record = Record.new(bam1, header)
+        yield record
+      end
+      self
+    end
   end
 end

data/lib/hts/bcf/format.rb

@@ -8,22 +8,22 @@ module HTS
         @p1 = FFI::MemoryPointer.new(:pointer) # FIXME: naming
       end
 
-      # For compatibility with htslib.cr.
+      # For compatibility with HTS.cr.
       def get_int(key)
         get(key, :int)
       end
 
-      # For compatibility with htslib.cr.
+      # For compatibility with HTS.cr.
       def get_float(key)
         get(key, :float)
       end
 
-      # For compatibility with htslib.cr.
+      # For compatibility with HTS.cr.
       def get_flag(key)
         get(key, :flag)
       end
 
-      # For compatibility with htslib.cr.
+      # For compatibility with HTS.cr.
       def get_string(key)
         get(key, :string)
       end

data/lib/hts/bcf/header.rb

@@ -1,11 +1,24 @@
 # frozen_string_literal: true
 
+require_relative "header_record"
+
 module HTS
   class Bcf < Hts
     # A class for working with VCF records.
+    # NOTE: This class has a lot of methods that are not stable.
+    # The method names and the number of arguments may change in the future.
     class Header
-      def initialize(hts_file)
-        @bcf_hdr = LibHTS.bcf_hdr_read(hts_file)
+      def initialize(arg = nil)
+        case arg
+        when LibHTS::HtsFile
+          @bcf_hdr = LibHTS.bcf_hdr_read(arg)
+        when LibHTS::BcfHdr
+          @bcf_hdr = arg
+        when nil
+          @bcf_hdr = LibHTS.bcf_hdr_init("w")
+        else
+          raise TypeError, "Invalid argument"
+        end
       end
 
       def struct
@@ -20,6 +33,10 @@ module HTS
         LibHTS.bcf_hdr_get_version(@bcf_hdr)
       end
 
+      def set_version(version)
+        LibHTS.bcf_hdr_set_version(@bcf_hdr, version)
+      end
+
       def nsamples
         LibHTS.bcf_hdr_nsamples(@bcf_hdr)
       end
@@ -31,6 +48,45 @@ module HTS
           .map(&:read_string)
       end
 
+      def add_sample(sample, sync: true)
+        LibHTS.bcf_hdr_add_sample(@bcf_hdr, sample)
+        self.sync if sync
+      end
+
+      def merge(hdr)
+        LibHTS.bcf_hdr_merge(@bcf_hdr, hdr.struct)
+      end
+
+      def sync
+        LibHTS.bcf_hdr_sync(@bcf_hdr)
+      end
+
+      def read_bcf(fname)
+        LibHTS.bcf_hdr_set(@bcf_hdr, fname)
+      end
+
+      def append(line)
+        LibHTS.bcf_hdr_append(@bcf_hdr, line)
+      end
+
+      def delete(bcf_hl_type, key) # FIXME
+        type = bcf_hl_type_to_int(bcf_hl_type)
+        LibHTS.bcf_hdr_remove(@bcf_hdr, type, key)
+      end
+
+      def get_hrec(bcf_hl_type, key, value, str_class = nil)
+        type = bcf_hl_type_to_int(bcf_hl_type)
+        hrec = LibHTS.bcf_hdr_get_hrec(@bcf_hdr, type, key, value, str_class)
+        HeaderRecord.new(hrec)
+      end
+
+      def seqnames
+        n = FFI::MemoryPointer.new(:int)
+        names = LibHTS.bcf_hdr_seqnames(@bcf_hdr, n)
+        names.read_array_of_pointer(n.read_int)
+             .map(&:read_string)
+      end
+
       def to_s
         kstr = LibHTS::KString.new
         raise "Failed to get header string" unless LibHTS.bcf_hdr_format(@bcf_hdr, 0, kstr)
@@ -40,6 +96,27 @@ module HTS
 
       private
 
+      def bcf_hl_type_to_int(bcf_hl_type)
+        return bcf_hl_type if bcf_hl_type.is_a?(Integer)
+
+        case bcf_hl_type.to_s.upcase
+        when "FILTER", "FIL"
+          LibHTS::BCF_HL_FLT
+        when "INFO"
+          LibHTS::BCF_HL_INFO
+        when "FORMAT", "FMT"
+          LibHTS::BCF_HL_FMT
+        when "CONTIG", "CTG"
+          LibHTS::BCF_HL_CTG
+        when "STRUCTURED", "STR"
+          LibHTS::BCF_HL_STR
+        when "GENOTYPE", "GEN"
+          LibHTS::BCF_HL_GEN
+        else
+          raise TypeError, "Invalid argument"
+        end
+      end
+
       def initialize_copy(orig)
         @bcf_hdr = LibHTS.bcf_hdr_dup(orig.struct)
       end

data/lib/hts/bcf/header_record.rb

@@ -3,9 +3,43 @@
 module HTS
   class Bcf < Hts
     class HeaderRecord
-      def initialize
+      def initialize(arg = nil)
+        case arg
+        when LibHTS::BcfHrec
+          @bcf_hrec = arg
+        else
+          raise TypeError, "Invalid argument"
+        end
+      end
+
+      def struct
         @bcf_hrec
       end
+
+      def add_key(key)
+        LibHTS.bcf_hrec_add_key(@bcf_hrec, key, key.length)
+      end
+
+      def set_value(i, val, quote: true)
+        is_quoted = quote ? 1 : 0
+        LibHTS.bcf_hrec_set_val(@bcf_hrec, i, val, val.length, is_quoted)
+      end
+
+      def find_key(key)
+        LibHTS.bcf_hrec_find_key(@bcf_hrec, key)
+      end
+
+      def to_s
+        kstr = LibHTS::KString.new
+        LibHTS.bcf_hrec_format(@bcf_hrec, kstr)
+        kstr[:s]
+      end
+
+      private
+
+      def initialize_copy(orig)
+        @bcf_hrec = LibHTS.bcf_hrec_dup(orig.struct)
+      end
     end
   end
 end

data/lib/hts/bcf/info.rb

@@ -9,22 +9,22 @@ module HTS
         @p1 = FFI::MemoryPointer.new(:pointer) # FIXME: naming
       end
 
-      # For compatibility with htslib.cr.
+      # For compatibility with HTS.cr.
       def get_int(key)
         get(key, :int)
       end
 
-      # For compatibility with htslib.cr.
+      # For compatibility with HTS.cr.
       def get_float(key)
         get(key, :float)
       end
 
-      # For compatibility with htslib.cr.
+      # For compatibility with HTS.cr.
       def get_string(key)
         get(key, :string)
       end
 
-      # For compatibility with htslib.cr.
+      # For compatibility with HTS.cr.
       def get_flag(key)
         get(key, :flag)
       end

data/lib/hts/bcf.rb

@@ -52,7 +52,6 @@ module HTS
       build_index(index) if build_index
       @idx = load_index(index)
       @start_position = tell
-      super # do nothing
     end
 
     def build_index(index_name = nil, min_shift: 14)
@@ -85,11 +84,16 @@ module HTS
       !@idx.null?
     end
 
-    def write_header
+    def close
+      LibHTS.hts_idx_destroy(@idx) unless @idx&.null?
+      @idx = nil
+      super
+    end
+
+    def write_header(header)
       check_closed
 
       @header = header.dup
-      LibHTS.hts_set_fai_filename(header, @file_name)
       LibHTS.bcf_hdr_write(@hts_file, header)
     end
 
@@ -97,7 +101,8 @@ module HTS
       check_closed
 
       var_dup = var.dup
-      LibHTS.bcf_write(@hts_file, header, var_dup) > 0 || raise
+      r = LibHTS.bcf_write(@hts_file, header, var_dup)
+      raise "Failed to write record" if r < 0
     end
 
     # Close the current file.
@@ -122,29 +127,6 @@ module HTS
       end
     end
 
-    private def each_record_copy
-      check_closed
-
-      return to_enum(__method__) unless block_given?
-
-      while LibHTS.bcf_read(@hts_file, header, bcf1 = LibHTS.bcf_init) != -1
-        record = Record.new(bcf1, header)
-        yield record
-      end
-      self
-    end
-
-    private def each_record_reuse
-      check_closed
-
-      return to_enum(__method__) unless block_given?
-
-      bcf1 = LibHTS.bcf_init
-      record = Record.new(bcf1, header)
-      yield record while LibHTS.bcf_read(@hts_file, header, bcf1) != -1
-      self
-    end
-
     def query(...)
       querys(...) # Fixme
     end
@@ -166,55 +148,6 @@ module HTS
     # private def queryi_reuse
     # end
 
-    private def querys_copy(region)
-      check_closed
-
-      raise "query is only available for BCF files" unless file_format == "bcf"
-      raise "Index file is required to call the query method." unless index_loaded?
-      return to_enum(__method__, region) unless block_given?
-
-      qitr = LibHTS.bcf_itr_querys(@idx, header, region)
-
-      begin
-        loop do
-          bcf1 = LibHTS.bcf_init
-          slen = LibHTS.hts_itr_next(@hts_file[:fp][:bgzf], qitr, bcf1, ::FFI::Pointer::NULL)
-          break if slen == -1
-          raise if slen < -1
-
-          yield Record.new(bcf1, header)
-        end
-      ensure
-        LibHTS.bcf_itr_destroy(qitr)
-      end
-      self
-    end
-
-    private def querys_reuse(region)
-      check_closed
-
-      raise "query is only available for BCF files" unless file_format == "bcf"
-      raise "Index file is required to call the query method." unless index_loaded?
-      return to_enum(__method__, region) unless block_given?
-
-      qitr = LibHTS.bcf_itr_querys(@idx, header, region)
-
-      bcf1 = LibHTS.bcf_init
-      record = Record.new(bcf1, header)
-      begin
-        loop do
-          slen = LibHTS.hts_itr_next(@hts_file[:fp][:bgzf], qitr, bcf1, ::FFI::Pointer::NULL)
-          break if slen == -1
-          raise if slen < -1
-
-          yield record
-        end
-      ensure
-        LibHTS.bcf_itr_destroy(qitr)
-      end
-      self
-    end
-
     # @!macro [attach] define_getter
     #   @method $1
     #   Get $1 array
@@ -285,5 +218,81 @@ module HTS
         yield r.format(key)
       end
     end
+
+    private
+
+    def querys_reuse(region)
+      check_closed
+
+      raise "query is only available for BCF files" unless file_format == "bcf"
+      raise "Index file is required to call the query method." unless index_loaded?
+      return to_enum(__method__, region) unless block_given?
+
+      qiter = LibHTS.bcf_itr_querys(@idx, header, region)
+      raise "Failed to query region #{region}" if qiter.null?
+
+      bcf1 = LibHTS.bcf_init
+      record = Record.new(bcf1, header)
+      begin
+        loop do
+          slen = LibHTS.hts_itr_next(@hts_file[:fp][:bgzf], qiter, bcf1, ::FFI::Pointer::NULL)
+          break if slen == -1
+          raise if slen < -1
+
+          yield record
+        end
+      ensure
+        LibHTS.bcf_itr_destroy(qiter)
+      end
+      self
+    end
+
+    def querys_copy(region)
+      check_closed
+
+      raise "query is only available for BCF files" unless file_format == "bcf"
+      raise "Index file is required to call the query method." unless index_loaded?
+      return to_enum(__method__, region) unless block_given?
+
+      qiter = LibHTS.bcf_itr_querys(@idx, header, region)
+      raise "Failed to query region #{region}" if qiter.null?
+
+      begin
+        loop do
+          bcf1 = LibHTS.bcf_init
+          slen = LibHTS.hts_itr_next(@hts_file[:fp][:bgzf], qiter, bcf1, ::FFI::Pointer::NULL)
+          break if slen == -1
+          raise if slen < -1
+
+          yield Record.new(bcf1, header)
+        end
+      ensure
+        LibHTS.bcf_itr_destroy(qiter)
+      end
+      self
+    end
+
+    def each_record_reuse
+      check_closed
+
+      return to_enum(__method__) unless block_given?
+
+      bcf1 = LibHTS.bcf_init
+      record = Record.new(bcf1, header)
+      yield record while LibHTS.bcf_read(@hts_file, header, bcf1) != -1
+      self
+    end
+
+    def each_record_copy
+      check_closed
+
+      return to_enum(__method__) unless block_given?
+
+      while LibHTS.bcf_read(@hts_file, header, bcf1 = LibHTS.bcf_init) != -1
+        record = Record.new(bcf1, header)
+        yield record
+      end
+      self
+    end
   end
 end

data/lib/hts/hts.rb

@@ -31,8 +31,8 @@ module HTS
       end
     end
 
-    def initialize(*args)
-      # do nothing
+    def initialize(*_args)
+      raise TypeError, "Can't make instance of HTS abstract class"
     end
 
     def struct